Diagnostic accuracy of preoperative ultrasonography in predicting contralateral inguinal hernia in children: a systematic review and meta-analysis

Inguinal hernia is the most common surgical condition in childhood and presents in approximately 0.8–5% of term children and more than 30% in premature born children [1, 2]. At the time of surgery, the hernia is unilateral in 80% of children. However, 7–15% of children develop a hernia on the opposite side after unilateral hernia repair, a metachronous contralateral inguinal hernia (MCIH), with the highest risk in children under the age of 6 months or with initial left-sided hernia [3, 4]. Medical history and physical examination of the groin are not adequate to detect a hidden (or asymptomatic) patent processus vaginalis (PV) [5‐7], which is likely to develop into an MCIH. MCIH necessitates a second operation and anaesthesia. The repeated use of anaesthetics possibly affect the development of children’s brains [8]. Preventive strategies have been proposed since the 1950s and exploration of the contralateral groin during unilateral hernia repair is frequently used to detect contralateral patent processus vaginalis (CPPV). However, contralateral exploration carries risks of operative complications (e.g. wound infection, haematoma or testicular atrophy) and is unnecessary if the processus vaginalis is closed [9, 10]. Therefore, management of the contralateral groin remains controversial.

If contralateral hernia development could be predicted prior to symptomatic hernia repair, individual recommendations regarding contralateral exploration can be made more consciously. Accurate preoperative evaluation of the contralateral inguinal canal could prevent unnecessary contralateral exploration and might reduce the incidence of children developing MCIH and obviating the need for a second operation and anaesthesia.

Previous studies show that preoperative ultrasonography has high accuracy for detecting both symptomatic and asymptomatic inguinal hernias in children, as the sensitivity for diagnosing CPPV is 98% and 97%, respectively [6, 11, 12]. False negative rates of ultrasonography-based detection of CPPV are difficult to determine since most studies do not evaluate the contralateral inguinal canal intra- or postoperatively [13].

The aim of this systematic review and meta-analysis is to assess the diagnostic accuracy of preoperative ultrasonography of the contralateral groin to detect CPPV and its potential value to predict development of MCIH in children with unilateral inguinal hernia.

In this review, we aimed to assess the diagnostic accuracy of preoperative ultrasonography of the contralateral groin to detect CPPV and its potential value to predict MCIH in children with unilateral inguinal hernia.

Seven studies were included in this meta-analysis. They were subdivided into two groups: ‘complete’ and ‘incomplete’ cases. Studies referred to as complete cases reported both the positive and negative index test results of contralateral exploration and showed high pooled sensitivity (93%) and moderate pooled specificity (88%). Incomplete cases reported only the positive index test results of contralateral exploration and the negative index test results of the findings of clinical follow-up, and show lower pooled sensitivity (86%) though higher pooled specificity (98%). AUC implies that diagnostic performance of preoperative ultrasonography is very good at detecting CPPV in complete (0.984) and incomplete cases (0.996). These results indicate that preoperative ultrasonography is less likely to result in a high number of false negatives in the complete cases. In the incomplete cases, the false positive rate (FPR) is low (2.1%) compared to the FPR of complete cases (11.5%). Negative preoperative ultrasonographic test results yield positive perioperative findings during contralateral exploration more often compared to the amount of clinically apparent hernias that actually develop postoperatively. These results suggest that patency of the contralateral PV is much more often detected during contralateral exploration than an MCIH actually develops during follow-up.

It is reported that at first presentation, 40% of the children diagnosed with a unilateral inguinal hernia also have a patent PV on the contralateral side [24‐26]. However, the process of obliteration of a patent PV continues in the first few months of life, causing a successive decrease in the PV patency incidence and thus the incidence of positive surgical findings with advancing age. A patent PV during contralateral exploration will therefore not necessarily develop into a clinically relevant inguinal hernia: the estimated childhood risk of developing a clinically apparent inguinal hernia in case of the presence of a patent PV is only between 25% and 50% [24]; however, it considerably increases the risk for MCIH development, and risk factors to predict whether a CPPV actually develops into an MCIH or not need to be identified. This corresponds to the results of Kaneda et al., who related both the positive and negative ultrasonographic test results to clinical follow-up and reported a high ‘false positive rate’ of 75% (27/36). In contrast to the other studies included in this review, the number of true positives is extremely low (Tables 3, 4 and 5), implying that only 25% of the CPPVs actually developed into a clinically apparent MCIH. Prevalence of CPPV as determined by preoperative ultrasonography was 34.3%, though after clinical follow-up the prevalence of MCIH decreased to 10.5% [13]. As the index test results are considered positive on the basis of the presence of CPPV, regardless whether it actually develops into a clinically apparent MCIH or not, one can assume that preoperative ultrasonography overestimates MCIH prevalence (Tables 2 and 3).

The preoperative major mean CPPV axis as determined using ultrasound was significantly wider in patients with contralateral inguinal hernia compared to patients without contralateral hernia. Unfortunately, there was great variability in the diameter of the contralateral WLIR among the included studies. The mean WLIR in children diagnosed with contralateral hernia or patent PV was 2.70 ± 1.17 mm, 6.8 ± 1.3 mm or 9.0 ± 1.9 mm [13, 20], whereas in other studies children with a mean WLIR of 1.34 ± 1.26 mm, 3.5 ± 0.4 mm and less than 6.0 mm were diagnosed as having no patent PV [13, 19, 20]. Previously, Erez et al. prospectively compared the sonographic dimensions of the symptomatic, unilateral inguinal canal with the surgical findings in 642 children undergoing inguinal hernia repair and perioperatively found a PPV, ‘full hernia’ and ‘large hernia’ when a mean groin width of 4.9 ± 1.1 mm, 7.2 ± 2.0 mm and 12.8 ± 3.6 mm was ultrasonically detected [27].

One of the difficulties interpreting our study results is that the diagnostic ultrasonographic criteria to detect CPPV differ among the included studies (Table 2). First, some studies differentiate between CPPV and contralateral inguinal hernia, while others consider CPPV as potential hernia sac and include them in the category ‘contralateral hernia’ [5, 14, 23]. Thus, unlike others, Erez et al. used the hypoechoic width at the middle of the inguinal canal as the point of reference, instead of e.g. the maximum diameter or width of the inguinal canal at the internal ring [27]. A CPPV or contralateral hernia is further considered as the patency of the PV, as the presence of fluid, bowel loops or omentum in an open PV, or as the diameter of the internal ring exceeds a certain value (e.g. ≥ 4 or 4.5 mm). This shows the necessity of establishing more adequate homogeneous ultrasonographic criteria to accurately detect CPPV [28]. The results should be interpreted with caution because of high heterogeneity. Most studies included in our meta-analysis (n = 7) only explored the contralateral groin if preoperative ultrasonography yielded positive test results. This indicates a higher probability of the presence of CPPV during surgery, which could lead to an overestimation of the index test results. False negative rates of ultrasonography (FNR) for CPPV and FNR regarding the occurrence of MCIH might also be imprecise, since only four studies surgically explored all contralateral groins irrespective of the preoperative ultrasonographic test results [14‐17] and only four studies have reported postoperative outcomes among patients who were not explored (Tables 1 and 2) [13, 18‐20]. The use of different cut-off values to detect CPPV in the included studies might cause threshold effects to occur, also leading to heterogeneity of the results. Unfortunately, this meta-analysis did not provide enough data to explore the effects of differences in study characteristics on the estimates of the diagnostic test accuracy.

In conclusion, preoperative ultrasonography of the contralateral groin to detect CPPV seems promising, but may result in an overestimation of MCIH prevalence. More adequate, unequivocal ultrasonographic criteria are mandatory for proper diagnosis of CPPV and careful consideration should be given to what the appropriate reference standard is to establish accurate diagnosis of MCIH. Finally, risk factors need to be identified that can predict whether a CPPV actually develops into a clinically relevant MCIH or not.